Could data help the dairy industry recover from the pandemic?

Lucy Saddler opens the lid on DairyMine, a research project that uses multiple datasets to support the dairy industry’s recovery from the pandemic.

With hospitality settings at a standstill from April through to May 2020 during the UK’s first lockdown, it is estimated that the financial impact felt by dairy farmers across the country was £20 million.

Even before the financial impact of the pandemic, dairy farms operated with high overheads and tight profit margins. The annual running cost for an average sized dairy farm is approximately £450,000 compared to an equivalent cost of £150,000 for a beef farm. This means efficiency key to profitable and sustainable dairy farming.

Dr Gregor Gorjanc, the principal investigator of DairyMine, explains how agriculture was impacted by Covid-19 in a variety of ways.

“In dairy production, some farmers actually had to throw away milk,” he says.

“It was a big shame. DairyMine is a way – not to address that specific problem, but to help the dairy sector emerge from the pandemic stronger by adopting new data driven and innovative breeding methods.”

With the aim of combining on-farm production and behavioural data with genomic data, Dr Gorjanc applied for funding through the Data Driven Innovation (DDI) programme’s Beacon funding call, which distributed funding from the Scottish Funding Council to data-driven projects to accelerate Scotland’s recovery from the pandemic.


The Roslin Institute

Dr Gorjanc is a Chancellor’s Fellow in Data-Driven Innovation for Agri-Tech and, alongside other genetic scientists on the DairyMine project, has drawn on the Roslin Institute’s history of data-driven breeding initiatives.

He explains: “The Roslin Institute has been involved in this line of [genomic] research for decades. The Institute is a world leader in the development of algorithms for breeding schemes and in thinking about how to record and use breeding data in the most optimal way. There have been iterations and iterations… decades of research in this space.”

As a result, the volume of agricultural data is increasing and the tools that scientists have at their disposal to explore this data are advancing. “The next step,” Dr Gorjanc adds, “is how do we get practical knowledge out of all that data and how do we this adopt this knowledge for animal management and breeding.”

There is already a national database available for these types of datasets run by the Agriculture and Horticulture Development Board and the Edinburgh Genetic Evaluation Service (EGENES) at Scotland’s Rural College (SRUC).

But currently these systems don’t fully exploit real-time and high frequency data from farms.

In bringing together on-farm production and behavioural data with genomic data, DairyMine seeks to use innovative real-time and high frequency data. This will come from a network of satellite farms monitored by Agri-EPI Centre – one of four Agritech centres established by the UK government, who have partnered with Dr Gorjanc and his team.

Dr Gorjanc uses the example of milk yield to illustrate the possibilities that real-time data can unlock for farmers.

“Obviously, we want cows who produce more milk. But if milk yield is measured just once or twice a month like it has been in the past, I can’t really see what’s happening between the months and within the month. Once you have continuous data you can start to find more nuanced patterns.”


How does DairyMine capitalise on the new datasets?

The basic concept behind DairyMine is to collect on-farm behavioural data, bring it together with genomic data and then analyse both datasets in tandem.

To collect this behavioural data, the team is drawing on Agri-EPI’s network of satellite farms, where various data recording devices and new wearable devices are being piloted on cattle.

“They are testing all sorts of different devices on these farms, some more established and some very novel,” explains Dr Gorjanc. “For example, some farms are already using milking robots. It’s not a human, but it’s an actual robot milking the cow. The robot will be automatically recording how much milk was produced and at what time of the day.”

Other devices used to collect data include what Dr Gorjanc describes as a “Fitbit for cows” that tracks a cow’s movement patterns in order to predict the cow’s reproductive status and rumination time.

Whilst the production and behavioural data is generated continually and collected on-farm, the genomic data is generated in a laboratory.

“We looked for variation at about 100,000 positions in the cow’s DNA. When you look at the DNA in any one position, you can see two variants. Let’s call one variant A and another variant B and so an individual cow can have a combination of AA, AB, or BB depending on what variant she had inherited from her father and her mother,” Dr Gorjanc explains.

The DairyMine team layer both the production and behavioural data from the farm and the genomic data to create a prediction equation. This is where Dr Gorjanc’s expertise lies.

The prediction equation allows them to surmise what behavioural traits a specific DNA sequence might be associated with, and it is this equation that will help inform breeding processes.

If a cow has a DNA sequence that corresponds to desirable traits such as a high milk yield or a good milk quality, dairy farmers would be more likely to breed this cow.

Dr Gorjanc explained that it is possible to get DNA from a newborn calf very quickly, thus allowing the farmer to use the prediction equation to ascertain how genetically good or bad that calf is for a range of traits, long before these traits are expressed.


Challenges ahead

Every project encounters challenges and Dr Gorjanc is open about those the DairyMine team encountered.

“One of the problems was that different farms have different recording devices,” he says.

Whilst some farms have many recording devices, others don’t use any of these technologies. It was specifically this variation that posed a big challenge for the team.

To grow DairyMine and expand its capabilities, the team will need to overcome these hurdles. But Dr Gorjanc is optimistic that this can be managed: “Where there is a challenge, there is an opportunity,” he concludes.


The project is funded by the Data-Driven Innovation initiative  (DDI),  delivered by the University of Edinburgh and Heriot-Watt University for the Edinburgh and South East Scotland City Region Deal. DDI is an innovation network helping organisations tackle challenges for industry and society by doing data right to support Edinburgh in its ambition to become the data capital of Europe.

3 rams eating grass

Read the latest Case studies

Photo of electric vehicle charging

Clearing the road ahead for the charging point market

It has been estimated the UK will need to increase its number of electric vehicle…

Photo of the team members of Carbon Glance

Carbon pricing solutions – at a Glance

Companies are increasingly exposed to regulations that price greenhouse gas emissions and Carbon Glance has…

Photo of NLP Natural Language Processing cognitive computing technology

Mining text unveils patterns in the brain

Spotting patterns in health across a population is a complex task, but sophisticated language technology…